OBJECTIVE: The human leukocyte antigen (HLA)-DRB1*13 allele frequency is lower in women with cervical carcinoma than in the general population, suggesting that this allele could exert a protective effect against progression of cervical intraepithelial neoplasia (CIN) associated with human papillomaviruses (HPV). To test this hypothesis, we designed a prospective study of low-grade CIN (CIN1) and analyzed the probability of regression of these lesions according to HLA-DR and HPV status.
METHODS: The study sample was composed of 86 women with CIN1 who agreed to regular colposcopic follow-up and no immediate treatment. Biopsy specimens were taken under colposcopy for histology and for the determination of HPV and HLA status. Cases were classified into 3 groups: CIN1 regression, persistence for at least 12 months, or progression to CIN2 or 3.
RESULTS: The rate of spontaneous regression (95% confidence interval) at 24 months was 51.6% (39–61.6%) overall compared with 34.7% (13.4–50.8%) in HPV16/18 positive cases and 59.9% (43.7–71.4%) in HPV16/18-negative cases (P = .051). The rate of regression was 71.8% (40.8–86.5%) in patients with HLA-DRB1*13 and 45.9% (31.5–57.2%) in patients with other genotypes (P = .03). Regression reached 90.5% (38.9–98.5%) at 18 months in DRB1*13 patients with HPV16/18-negative–associated CIN (15.1% of the cases). In multivariable analysis, HLA-DRB1*13 allele and HPV16/18-negative status were independently associated with an increased probability for regression (adjusted hazard ratio 2.1 [1.0–4.1] and 2.5 [1.2–5.4], respectively).
CONCLUSION: A subset of approximately 15% of CIN1 highly likely to show spontaneous regression can be defined using 2 biologic parameters that characterize the viral causative agent and the host.
LEVEL OF EVIDENCE: II-2
Approximately 90% of CIN1 occurring in HLA-DRB1*13 patients and associated with HPV sequences other than HPV16 or 18 show spontaneous short-term regression.
From the Departments of *Pathology, ‡Biostatistics, and §Physiopathology, Institut Curie; †Laboratory of Dr. René Cartier; ¶Laboratory of Immunology and Institut National de la Santé et de la Recherche Médicale U396, Hôpital Saint-Louis, Paris, France.
Address reprint requests to: Xavier Sastre-Garau, MD, PhD, Laboratoire de Pathologie, Institut Curie, 26 rue d'Ulm, 75231, Paris Cedex, France; e-mail: email@example.com.
Received April 8, 2004. Received in revised form June 11, 2004. Accepted June 24, 2004.
Most neoplasias of the uterine cervix are associated with specific types of human papillomaviruses (HPV). It is currently believed that invasive carcinoma is preceded by a long period of cervical intraepithelial neoplasia (CIN), a heterogeneous group of diseases classified as CIN1, CIN2, or CIN3 according to the severity of their histologic abnormalities.1 The most frequent lesion is CIN1, particularly in young women, and may follow different courses without treatment: spontaneous regression occurs in 50–60% of patients, whereas persistence or progression to CIN2 or 3 is observed in the rest.2 The oncogenicity of associated HPV types has been reported to influence the disease outcome,3 but little is known concerning the role of host-linked characteristics, such as immunologic factors, also likely to be involved in the natural history of these lesions. Cell-mediated immunity has been reported to be involved in the host response against specific HPV types4 and may thus account, at least in part, for regression of CIN. In HPV16-associated disease, clearance of the infection and CIN regression have been found to be associated with lymphoproliferative responses to specific HPV16 E6 and E7 epitopes.4 In invasive cancer, a frequent down-regulation of human leukocyte antigen (HLA) class I alleles has been observed,5 which may account for the lack of the immune response against HPV-positive carcinoma cells. Experimental and clinical studies have also suggested that immunogenetic determinants related to HLA class II genotype may play a part in tumor progression. In rabbit tumor models induced by the Shope papillomavirus, regression and malignant conversion of papillomas were found to be linked to major histocompatibility complex class II genes.6 In humans, several studies showed positive associations between invasive cervical cancer and DQB1*301,7,8 DQB1*1501,9 or DQB1*11017,10 HLA class II alleles, whereas no significant association was found in another group of patients.11 We and others, however, have consistently reported a negative association of the DRB1*13 allele with cervical cancer in patients from different geographic areas, ie, Europe,12,13 Africa,10 or North America.7 DRB1*13 or related genes could thus exert a protective effect against HPV-associated lesions. To test the hypothesis that regression of CIN could be associated with the DRB1*13 allele, we carried out a prospective study based on the colposcopy follow-up of women with CIN1 and compared the probability of regression of the lesions according to their HPV and HLA-DR status.
MATERIALS AND METHODS
The cohort analyzed was composed of 86 consecutive patients with CIN1, referred between May 1990 and March 1995 for a colposcopy due to an abnormal Pap test and who agreed to regular colposcopic follow-up and no immediate treatment. Colposcopies and histologic analyses were performed by a single expert gynecologic pathologist (I.C.) in private practice. To optimize endocervical examination and visualization of the squamous-columnar junction, all colposcopies were carried out between the 8th and 13th day of the menstrual cycle after 1 week of treatment with 50 μg ethinyl estradiol (E2) per day. Two contiguous biopsy specimens of the lesion were taken under colposcopic control, one fixed in Bouin's solution for histologic analysis and the other one frozen for HPV and HLA typing. Only cases with adequate colposcopy (fully visible squamous-columnar junction) and CIN1 were included in the study. The follow-up protocol included a second colposcopy with biopsy 10–12 months after the first diagnosis of CIN1 and at yearly intervals thereafter. Some patients consulted earlier for personal convenience. Cases were classified in 3 categories: regression when no more evidence of CIN was found at colposcopic and histologic analyses performed during follow-up, persistence when the lesion remained a CIN1 over a period of at least 12 months, and progression when the subsequent histology was CIN2 or 3. The study received institutional approval by the Gynecologic Cancer Study Group.
HPV typing was performed as previously reported.13 Cases positive by polymerase chain reaction technique using consensus primers (GP5+/GP6+) only were referred as HPVX. The DNA extracted from tumor tissue specimens for HPV typing was also used for HLA typing. The typing of DQA1, DQB1, and DRB1 was performed by polymerase chain reaction specific sequence oligonucleotide (SSO) reverse dot blot (INNO-LiPA, Innogenetics, Ghent, Belgium).
For statistical analysis, χ2 or Fisher exact tests were used when appropriate to compare proportions. The Kruskal-Wallis test was used when necessary to compare means. We estimated the risk of regression at 24 months using the Kaplan-Meier product-limit method. Log rank tests were used for comparing the risk of regression between the subgroups of potential risk factors. The Cox model was used for multivariate analysis. All tests were 2-tailed, and differences were considered significant at P < .05. The analysis was carried out using the S-PLUS 2000 software (Insightful Inc., Seattle, WA).
Mean age of the patients was 31 (range 18–70) years. During a median follow-up of 24 months, 41 of 86 cases showed regression, whereas 27 lesions remained CIN1, and 18 progressed to CIN2 or 3 (Table 1). In 69 (80%) cases HPV DNA sequences were detected (Table 1). Sequences detected corresponded to HPV16 (26 cases), HPV33 (7 cases), HPV31 (5 cases), HPV51 (3 cases), HPV6, 18, 35, 51 (2 cases each), and HPVX (20 cases). The HLA-DRB1*13 genotype was found in 20 (23%) cases (Table 1), of which 9 cases corresponded to DRB1*1301, 9 cases to DRB1*1302, and 2 cases to DRB1*1303.
Analysis of disease outcome showed that the overall rate of regression was 23.8% (95% confidence interval 14.1–32.4%) at 12 months and 51.6% (39–61.6%) at 24 months. This rate was 61.9% (44.8–73.7%) in women aged 30 years or younger and 36.8% (17.4–51.7%) in women aged older than 30 years (P = .03). Comparing the course of the disease to the HPV status, the rate of regression was 34.7% (13.4–50.8%) at 24 months in cases associated with HPV16/18 and 59.9% (43.7–71.4%) in the remaining cases (unadjusted hazard ratio [HR] = 2.0 [1.0–4.3], P = .051) (Fig. 1). Comparison between HLA-DR status and disease outcome showed that the 24-months rate of regression was 71.8% (40.8–86.5%) in patients with HLA-DRB1*13 and 45.9% (31.5–57.2%) in patients with other genotypes (unadjusted HR = 2.1 [1.1–4.0], P = .03) (Fig. 2). Eight of 9 patients with DRB1*1302 genotype showed regression.
The probability of regression was also analyzed according to HPV and HLA-DRB1*13 status. In the group of HLA-DRB1*13 patients with HPV16/18-negative–associated CIN1 (15.1% of the cases), the regression rate was 90.5% (38.9–98.5%) (Fig. 3). The lowest regression rate, 31.8% (6.9–50%), was found in patients with HLA-DRB1*13-negative allele and presenting CIN1 associated with HPV16/18. The other 2 groups showed 42.9% (0–69.9%) and 52.1% (33.9–65.3%) (Fig. 3) (P = .003). No significant difference was found when comparing mean ages among the 4 groups of patients.
We performed multivariable analysis using age, DRB1*13 status, and HPV16/18 status as covariates. The adjusted HR for regression within 24 months in HLA-DRB1*13 patients compared with patients with other alleles was 2.1 (1.0–4.1) (P < .04). The adjusted HR for patients with negative HPV16/18 status compared with those with positive HPV16/18 status was 2.5 (1.2–5.4) (P = .02). Age did not contribute significantly to the final model. No significant interaction was found between HPV status and DRB1*13 status. The DRB1*13 status adjusted on HPV16/18 status still played an independent role on regression. Furthermore, no significant association was observed between the course of CIN1 and the frequency of other alleles previously reported to present positive or negative associations with invasive cancer (DRB1*1001, DRB1*1101, DRB1*1501, DRB1*0301, or DQB1*0301, data not shown).
This prospective study shows that the course of CIN1 is related to both the associated-HPV type and the HLA-DR genotype of the patients. Spontaneous short-term regression was observed in more than 90% of CIN1 developed in HLA-DRB1*13 patients with HPV16/18-negative status, comprising 15% of our cases. The HLA-DRB1*13 genotype frequency was 23% in our CIN1 patients, a rate similar to that found in the general population, but significantly higher than that reported in patients with invasive carcinoma of the cervix from a variety of ethnic groups.7,9,10,13 This difference correlates well with the increased rate of regression of CIN1 in HLA-DRB1*13 patients observed in the present study. This association could be related to any DRB1*13 linked genes; however, in our study, the DRB1*1302 genotype was found to be more closely associated with disease outcome than the DRB1*1301/1303 genotypes and was not preferentially associated with any specific HLA-DQ alleles (data not shown). This suggests a role of the DRB1*1302 gene product in the immune response against HPV infection, but the involvement of other genes associated with this allele cannot be excluded.
We did not observe a predominance of specific HPV type in the HLA-DRB1*13 patients. Further analyses are necessary to determine whether this association could be due to the presentation of HPV peptides common to various viral types by the HLA-DRB1*13, leading to an efficient immune response. In HPV16/18-associated lesions, the positive DRB1*13 effect may be counterbalanced by the high oncogenicity of these virus types. Even in HLA-DRB1*13-negative patients associated with HPV16/18, the spontaneous rate of regression was still approximately 30%. The increased frequency of the DQB1*0301 allele, which has been reported in CIN1/214 and CIN3,15,16 was not seen in our patient population.
In conclusion, our analysis showed that regression of CIN1 was preferentially associated with the HLA-DRB1*13 genotype, particularly in cases with HPV types of intermediate oncogenicity. A CIN1 outcome could thus be predicted in a large proportion of cases by the assessment of viral and host-immunogenetic factors. Most CIN1 do not require specific treatment. The classical concept of progression from CIN1 to CIN2 or 3 is controversial,17 and it has been shown that high-grade CIN may develop de novo. Whether our results could be extended to the evolution of CIN2 or 3 is important. The proportion of DRB1*13 patients with HPV16/18 status-negative represented less than 2% of the population of invasive cancers analyzed in our previous study.13 In addition to lower DRB1*13 genotype frequency found in women with invasive cancer from various geographic origin, this strongly suggests that the natural history of high-grade CIN also depends, at least in part, on the HLA-DR and HPV status of the patients. Confirmation of this hypothesis would be of great practical clinical interest. Objective biologic criteria would permit the differentiation of CIN2 or 3 corresponding to potential precursors of invasive cancer requiring immediate ablative therapy from those cases in which treatment could be delayed. This would allow a substantial benefit in terms of cost and morbidity in the care of high-grade CIN, particularly in young women. Because 40% to 50% of CIN2 or 3 correspond to HPV16/18-negative lesions18,19 and the frequency of HLA-DRB1*13 is 22–29% in the general population,7,13 9–15% of the patients with de novo CIN2 or 3 would be expected to be found in the high regression group. The definition of the viral epitopes at the basis of the association reported here is of interest in the field of prophylactic and curative immunotherapy.
1. Richart RM. A modified terminology for cervical intraepithelial neoplasia. Obstet Gynecol 1990;75:131–3.
2. Östör AG. Natural history of cervical intraepithelial neoplasia: a critical review. Int J Gynecol Pathol 1993;12:186–92.
3. Schlecht NF, Platt RW, Duarte-Franco E, Costa MC, Sobrinho JP, Prado JC, et al. Human papillomavirus infection and time to progression and regression of cervical intraepithelial neoplasia. J Natl Cancer Inst 2003;95:1336–43.
4. Kadish AS, Ho GY, Burk RD, Wang Y, Romney SL, Ledwidge R, et al. Lymphoproliferative responses to human papillomavirus (HPV) type 16 proteins E6 and E7: outcome of HPV infection and associated neoplasia. J Natl Cancer Inst 1997;89:1285–93.
5. Keating PJ, Cromme FV, Duggan-Keen M, Snijders PJ, Walboomers JM, Hunter RD, et al. Frequency of down-regulation of individual HLA-A and -B alleles in cervical carcinomas in relation to TAP-1 expression. Br J Cancer 1995;72:405–11.
6. Han R, Breitburd F, Marche P, Orth G. Linkage of regression and malignant conversion of rabbit viral papillomas to MHC class II genes. Nature 1992;356:66–8.
7. Madeleine MM, Brumback B, Cushing-Haugen KL, Schwartz SM, Daling JR, Smith AG, et al. Human leukocyte antigen class II and cervical cancer risk: a population-based study. J Infect Dis 2002;186:1565–74.
8. Wank R, Thomssen C. High risk of squamous cell carcinoma of the cervix for women with HLA-DQw3. Nature 1991;352:723–5.
9. Apple RJ, Erlich HA, Klitz W, Manos MM, Becker TM, Wheeler CM. HLA DR-DQ associations with cervical carcinoma show papillomavirus-type specificity. Nat Genet 1994;6:157–62.
10. Lin P, Koutsky LA, Critchlow CW, Apple RJ, Hawes SE, Hughes JP, et al. HLA class II DR-DQ and increased risk of cervical cancer among Senegalese women. Cancer Epidemiol Biomarkers Prev 2001;10:1037–45.
11. Glew SS, Duggan-Keen M, Ghosh AK, Ivinson A, Sinnott P, Davidson J, et al. Lack of association of HLA polymorphisms with human papillomavirus-related cervical cancer. Hum Immunol 1993;37:157–64.
12. Apple RJ, Becker TM, Wheeler CM, Erlich HA. Comparison of human leukocyte antigen DR-DQ disease associations found with cervical dysplasia and invasive cervical carcinoma. J Natl Cancer Inst 1995;87:427–36.
13. Sastre-Garau X, Loste MN, Vincent-Salomon A, Favre M, Mouret E, de la Rochefordiere A, et al. Decreased frequency of HLA-DRB1 13 alleles in Frenchwomen with HPV-positive carcinoma of the cervix. Int J Cancer 1996;69:159–64.
14. Vandenvelde C, De Foor M, van Beers D. HLA-DOB1*03 and cervical intraepithelial neoplasia grades I-III. Lancet 1993;341:442.
15. David AL, Taylor GM, Gokhale D, Aplin JD, Seif MW, Tindall VR. HLA-DQB1*03 and cervical intraepithelial neoplasia type III. Lancet 1992;340:52.
16. Odunsi K, Terry G, Ho L, Bell J, Cuzick J, Ganesan TS. Susceptibility to human papillomavirus-associated cervical intra-epithelial neoplasia is determined by specific HLA DR-DQ alleles. Int J Cancer 1996;67:595–602.
17. Wright TC, Kurman RJ. A critical review of the morphologic classification systems of preinvasive lesions of the cervix: the scientific basis for shifting the paradigm. Papillomavirus Rep 1994;5:175–82.
18. Bergeron C, Barrasso R, Beaudenon S, Flamant P, Croissant O, Orth G. Human papillomaviruses associated with cervical intraepithelial neoplasia: great diversity and distinct distribution in low- and high-grade lesions. Am J Surg Pathol 1992;16:641–9.
© 2004 by The American College of Obstetricians and Gynecologists.
19. Lorincz AT, Reid R, Jenson AB, Greenberg MD, Lancaster W, Kurman RJ. Human papillomavirus infection of the cervix: relative risk associations of 15 common anogenital types. Obstet Gynecol 1992;79:328–37.